Notes from inside the work.
Short essays on agent architecture in financial services, the design trade-offs between conversational narrative and data privacy, and what it has actually been like to ship production AI software with both major agentic IDEs of 2025–2026.
Claude Code, Google Antigravity, and shipping production AI software in 2026.
Notes from shipping both halves of an AI agent platform with both major agentic IDEs of 2025–2026. What I learned, where each shines, and why I migrated.
MCP servers and the buy-side: what it changes, what it doesn't.
Reflections from publishing a production Model Context Protocol server for buy-side investment-management data — and what I'd tell another vendor before they shipped one.
The privacy trade-off no AI agent paper wants to discuss.
The honest two-part framing for AI agents on regulated data: schema-only at query construction, results sent under commercial privacy guarantees at narrative time. Why precision earns more trust than the simple claim.